Data summary

Raw Data Tally

We are missing 3 raw data files in total that we should have:

Note: subject 03-001 was not scanned at NIH site; subject 03-002 was not scanned at Penn.

Total visits per subjects

Scans at site per subject

Missing segmentations

Showing only segmentation types with at least 1 file missing.

Segmentation volumetrics

Site effects

Paired t-tests didnot find a significant difference.

## [[1]]
## 
##  Paired t-test
## 
## data:  JLF_WM by recon
## t = 0.0081621, df = 10, p-value = 0.9936
## alternative hypothesis: true mean difference is not equal to 0
## 95 percent confidence interval:
##  -0.2434550  0.2452452
## sample estimates:
## mean difference 
##    0.0008951038 
## 
## 
## [[2]]
## 
##  Paired t-test
## 
## data:  JLF_GM by recon
## t = -1.1724, df = 10, p-value = 0.2682
## alternative hypothesis: true mean difference is not equal to 0
## 95 percent confidence interval:
##  -0.29414720  0.09132281
## sample estimates:
## mean difference 
##      -0.1014122 
## 
## 
## [[3]]
## 
##  Paired t-test
## 
## data:  FIRST_thal by recon
## t = 0.28004, df = 10, p-value = 0.7852
## alternative hypothesis: true mean difference is not equal to 0
## 95 percent confidence interval:
##  -0.2326935  0.2995922
## sample estimates:
## mean difference 
##      0.03344938 
## 
## 
## [[4]]
## 
##  Paired t-test
## 
## data:  JLF_thal by recon
## t = -0.42861, df = 10, p-value = 0.6773
## alternative hypothesis: true mean difference is not equal to 0
## 95 percent confidence interval:
##  -0.3362662  0.2277673
## sample estimates:
## mean difference 
##     -0.05424947 
## 
## 
## [[5]]
## 
##  Paired t-test
## 
## data:  mimosa by recon
## t = 1.0647, df = 10, p-value = 0.312
## alternative hypothesis: true mean difference is not equal to 0
## 95 percent confidence interval:
##  -0.1516548  0.4292237
## sample estimates:
## mean difference 
##       0.1387844 
## 
## 
## [[6]]
## 
##  Paired t-test
## 
## data:  FAST_TBV by recon
## t = -2.3734, df = 10, p-value = 0.03905
## alternative hypothesis: true mean difference is not equal to 0
## 95 percent confidence interval:
##  -0.42142807 -0.01330599
## sample estimates:
## mean difference 
##       -0.217367

Permutation Testing

Gadgetron Volumes

pseudo-F Ratio statistics: Intersite Variability / Intrasite Variability

Distribution of permutated pseudo-F stats:

P-value of pseudo-F stat:

##     JLF_WM     JLF_GM FIRST_thal   JLF_thal     mimosa   FAST_TBV 
## 0.00609939 0.05309469 0.01519848 0.16568343 0.00289971 0.05419458

Onscanner Volumes

pseudo-F statistics: Intersite Variability / Intrasite Variability

Distribution of permutated pseudo-F stats:

P-value of pseudo-F stat:

##     JLF_WM     JLF_GM FIRST_thal   JLF_thal     mimosa   FAST_TBV 
## 0.00009999 0.00189981 0.07659234 0.00889911 0.05389461 0.00009999

Mixed Models

## $JLF_WM
## Linear mixed model fit by REML ['lmerMod']
## Formula: JLF_WM ~ (1 | ID) + (1 | site:ID)
##    Data: vol_df
## 
## REML criterion at convergence: 472.9
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -4.0446 -0.2112  0.1467  0.3164  2.1308 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  site:ID  (Intercept)  34.87    5.905  
##  ID       (Intercept) 502.56   22.418  
##  Residual              58.55    7.652  
## Number of obs: 61, groups:  site:ID, 31; ID, 11
## 
## Fixed effects:
##             Estimate Std. Error t value
## (Intercept)  510.142      6.918   73.75
## 
## $JLF_GM
## Linear mixed model fit by REML ['lmerMod']
## Formula: JLF_GM ~ (1 | ID) + (1 | site:ID)
##    Data: vol_df
## 
## REML criterion at convergence: 498.9
## 
## Scaled residuals: 
##      Min       1Q   Median       3Q      Max 
## -2.15565 -0.46933  0.02882  0.29009  2.96851 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  site:ID  (Intercept)  92.53    9.619  
##  ID       (Intercept) 971.22   31.164  
##  Residual              70.74    8.411  
## Number of obs: 61, groups:  site:ID, 31; ID, 11
## 
## Fixed effects:
##             Estimate Std. Error t value
## (Intercept)  712.946      9.622    74.1
## 
## $FIRST_thal
## Linear mixed model fit by REML ['lmerMod']
## Formula: FIRST_thal ~ (1 | ID) + (1 | site:ID)
##    Data: vol_df
## 
## REML criterion at convergence: 298.9
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -3.7565 -0.0293  0.1583  0.3316  1.9379 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  site:ID  (Intercept) 3.369    1.835   
##  ID       (Intercept) 1.833    1.354   
##  Residual             5.158    2.271   
## Number of obs: 60, groups:  site:ID, 31; ID, 11
## 
## Fixed effects:
##             Estimate Std. Error t value
## (Intercept)  12.9406     0.6042   21.42
## 
## $JLF_thal
## Linear mixed model fit by REML ['lmerMod']
## Formula: JLF_thal ~ (1 | ID) + (1 | site:ID)
##    Data: vol_df
## 
## REML criterion at convergence: 125.4
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -4.8226 -0.2199  0.0098  0.3410  1.3244 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  site:ID  (Intercept) 0.04114  0.2028  
##  ID       (Intercept) 1.28938  1.1355  
##  Residual             0.22288  0.4721  
## Number of obs: 61, groups:  site:ID, 31; ID, 11
## 
## Fixed effects:
##             Estimate Std. Error t value
## (Intercept)  13.5394     0.3499    38.7
## 
## $mimosa
## Linear mixed model fit by REML ['lmerMod']
## Formula: mimosa ~ (1 | ID) + (1 | site:ID)
##    Data: vol_df
## 
## REML criterion at convergence: 475.1
## 
## Scaled residuals: 
##      Min       1Q   Median       3Q      Max 
## -2.54186 -0.32830 -0.01482  0.47758  1.61987 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  site:ID  (Intercept) 84.74    9.206   
##  ID       (Intercept) 62.42    7.900   
##  Residual             86.85    9.319   
## Number of obs: 60, groups:  site:ID, 31; ID, 11
## 
## Fixed effects:
##             Estimate Std. Error t value
## (Intercept)   38.917      3.154   12.34
## 
## $FAST_TBV
## Linear mixed model fit by REML ['lmerMod']
## Formula: FAST_TBV ~ (1 | ID) + (1 | site:ID)
##    Data: vol_df
## 
## REML criterion at convergence: 628.6
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -4.5039 -0.1621  0.1276  0.4302  2.7070 
## 
## Random effects:
##  Groups   Name        Variance  Std.Dev. 
##  site:ID  (Intercept) 9.391e-11 9.691e-06
##  ID       (Intercept) 1.953e+03 4.419e+01
##  Residual             1.349e+03 3.673e+01
## Number of obs: 61, groups:  site:ID, 31; ID, 11
## 
## Fixed effects:
##             Estimate Std. Error t value
## (Intercept)  1028.74      14.16   72.64
## optimizer (nloptwrap) convergence code: 0 (OK)
## boundary (singular) fit: see help('isSingular')

ICC of Subject Random Effect

ICC of Site:Subject Random Effect

Mixed model Diagnostics

FAST TBV

## $summary
## Linear mixed model fit by REML ['lmerMod']
## Formula: FAST_TBV ~ (1 | ID) + (1 | site:ID)
##    Data: vol_df
## 
## REML criterion at convergence: 628.6
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -4.5039 -0.1621  0.1276  0.4302  2.7070 
## 
## Random effects:
##  Groups   Name        Variance  Std.Dev. 
##  site:ID  (Intercept) 9.391e-11 9.691e-06
##  ID       (Intercept) 1.953e+03 4.419e+01
##  Residual             1.349e+03 3.673e+01
## Number of obs: 61, groups:  site:ID, 31; ID, 11
## 
## Fixed effects:
##             Estimate Std. Error t value
## (Intercept)  1028.74      14.16   72.64
## optimizer (nloptwrap) convergence code: 0 (OK)
## boundary (singular) fit: see help('isSingular')
## 
## 
## $perf
## [1] NA
## 
## $shapiro.resid
## 
##  Shapiro-Wilk normality test
## 
## data:  residuals(model)
## W = 0.78891, p-value = 6.012e-08
## 
## 
## $qq.resid
## [1] 61 60
## 
## $shapiro.ID
## 
##  Shapiro-Wilk normality test
## 
## data:  coef(model)$ID[, 1]
## W = 0.91642, p-value = 0.29
## 
## 
## $`shapiro.site:ID`
## 
##  Shapiro-Wilk normality test
## 
## data:  coef(model)$`site:ID`[, 1]
## W = 0.90285, p-value = 0.008495
## 
## 
## $qq.ranef.ID
## [1] 6 8
## 
## $`qq.ranef.site:ID`
## [1] 24 25
## 
## $res.v.fit

JLF GM

## $summary
## Linear mixed model fit by REML ['lmerMod']
## Formula: JLF_GM ~ (1 | ID) + (1 | site:ID)
##    Data: vol_df
## 
## REML criterion at convergence: 498.9
## 
## Scaled residuals: 
##      Min       1Q   Median       3Q      Max 
## -2.15565 -0.46933  0.02882  0.29009  2.96851 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  site:ID  (Intercept)  92.53    9.619  
##  ID       (Intercept) 971.22   31.164  
##  Residual              70.74    8.411  
## Number of obs: 61, groups:  site:ID, 31; ID, 11
## 
## Fixed effects:
##             Estimate Std. Error t value
## (Intercept)  712.946      9.622    74.1
## 
## $perf
## # ICC by Group
## 
## Group   |   ICC
## ---------------
## site:ID | 0.082
## ID      | 0.856
## 
## $shapiro.resid
## 
##  Shapiro-Wilk normality test
## 
## data:  residuals(model)
## W = 0.9448, p-value = 0.008218
## 
## 
## $qq.resid
## [1] 7 9
## 
## $shapiro.ID
## 
##  Shapiro-Wilk normality test
## 
## data:  coef(model)$ID[, 1]
## W = 0.65902, p-value = 0.0001402
## 
## 
## $`shapiro.site:ID`
## 
##  Shapiro-Wilk normality test
## 
## data:  coef(model)$`site:ID`[, 1]
## W = 0.9114, p-value = 0.01404
## 
## 
## $qq.ranef.ID
## [1] 6 9
## 
## $`qq.ranef.site:ID`
## [1] 13  2
## 
## $res.v.fit

JLF WM

## $summary
## Linear mixed model fit by REML ['lmerMod']
## Formula: JLF_WM ~ (1 | ID) + (1 | site:ID)
##    Data: vol_df
## 
## REML criterion at convergence: 472.9
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -4.0446 -0.2112  0.1467  0.3164  2.1308 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  site:ID  (Intercept)  34.87    5.905  
##  ID       (Intercept) 502.56   22.418  
##  Residual              58.55    7.652  
## Number of obs: 61, groups:  site:ID, 31; ID, 11
## 
## Fixed effects:
##             Estimate Std. Error t value
## (Intercept)  510.142      6.918   73.75
## 
## $perf
## # ICC by Group
## 
## Group   |   ICC
## ---------------
## site:ID | 0.059
## ID      | 0.843
## 
## $shapiro.resid
## 
##  Shapiro-Wilk normality test
## 
## data:  residuals(model)
## W = 0.81664, p-value = 3.03e-07
## 
## 
## $qq.resid
## [1] 60 61
## 
## $shapiro.ID
## 
##  Shapiro-Wilk normality test
## 
## data:  coef(model)$ID[, 1]
## W = 0.95364, p-value = 0.6907
## 
## 
## $`shapiro.site:ID`
## 
##  Shapiro-Wilk normality test
## 
## data:  coef(model)$`site:ID`[, 1]
## W = 0.95343, p-value = 0.194
## 
## 
## $qq.ranef.ID
## [1] 6 1
## 
## $`qq.ranef.site:ID`
## [1] 31 21
## 
## $res.v.fit

FIRST thal

## $summary
## Linear mixed model fit by REML ['lmerMod']
## Formula: FIRST_thal ~ (1 | ID) + (1 | site:ID)
##    Data: vol_df
## 
## REML criterion at convergence: 298.9
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -3.7565 -0.0293  0.1583  0.3316  1.9379 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  site:ID  (Intercept) 3.369    1.835   
##  ID       (Intercept) 1.833    1.354   
##  Residual             5.158    2.271   
## Number of obs: 60, groups:  site:ID, 31; ID, 11
## 
## Fixed effects:
##             Estimate Std. Error t value
## (Intercept)  12.9406     0.6042   21.42
## 
## $perf
## # ICC by Group
## 
## Group   |   ICC
## ---------------
## site:ID | 0.325
## ID      | 0.177
## 
## $shapiro.resid
## 
##  Shapiro-Wilk normality test
## 
## data:  residuals(model)
## W = 0.75021, p-value = 9.414e-09
## 
## 
## $qq.resid
## 24 18 
## 23 17 
## 
## $shapiro.ID
## 
##  Shapiro-Wilk normality test
## 
## data:  coef(model)$ID[, 1]
## W = 0.92452, p-value = 0.3581
## 
## 
## $`shapiro.site:ID`
## 
##  Shapiro-Wilk normality test
## 
## data:  coef(model)$`site:ID`[, 1]
## W = 0.88576, p-value = 0.003239
## 
## 
## $qq.ranef.ID
## [1] 7 9
## 
## $`qq.ranef.site:ID`
## [1] 24 25
## 
## $res.v.fit

JLF thal

## $summary
## Linear mixed model fit by REML ['lmerMod']
## Formula: JLF_thal ~ (1 | ID) + (1 | site:ID)
##    Data: vol_df
## 
## REML criterion at convergence: 125.4
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -4.8226 -0.2199  0.0098  0.3410  1.3244 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  site:ID  (Intercept) 0.04114  0.2028  
##  ID       (Intercept) 1.28938  1.1355  
##  Residual             0.22288  0.4721  
## Number of obs: 61, groups:  site:ID, 31; ID, 11
## 
## Fixed effects:
##             Estimate Std. Error t value
## (Intercept)  13.5394     0.3499    38.7
## 
## $perf
## # ICC by Group
## 
## Group   |   ICC
## ---------------
## site:ID | 0.026
## ID      | 0.830
## 
## $shapiro.resid
## 
##  Shapiro-Wilk normality test
## 
## data:  residuals(model)
## W = 0.70624, p-value = 9.497e-10
## 
## 
## $qq.resid
## [1] 17 60
## 
## $shapiro.ID
## 
##  Shapiro-Wilk normality test
## 
## data:  coef(model)$ID[, 1]
## W = 0.95519, p-value = 0.7106
## 
## 
## $`shapiro.site:ID`
## 
##  Shapiro-Wilk normality test
## 
## data:  coef(model)$`site:ID`[, 1]
## W = 0.88676, p-value = 0.003423
## 
## 
## $qq.ranef.ID
## [1] 1 9
## 
## $`qq.ranef.site:ID`
## [1] 24 14
## 
## $res.v.fit

Mimosa

## $summary
## Linear mixed model fit by REML ['lmerMod']
## Formula: mimosa ~ (1 | ID) + (1 | site:ID)
##    Data: vol_df
## 
## REML criterion at convergence: 475.1
## 
## Scaled residuals: 
##      Min       1Q   Median       3Q      Max 
## -2.54186 -0.32830 -0.01482  0.47758  1.61987 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  site:ID  (Intercept) 84.74    9.206   
##  ID       (Intercept) 62.42    7.900   
##  Residual             86.85    9.319   
## Number of obs: 60, groups:  site:ID, 31; ID, 11
## 
## Fixed effects:
##             Estimate Std. Error t value
## (Intercept)   38.917      3.154   12.34
## 
## $perf
## # ICC by Group
## 
## Group   |   ICC
## ---------------
## site:ID | 0.362
## ID      | 0.267
## 
## $shapiro.resid
## 
##  Shapiro-Wilk normality test
## 
## data:  residuals(model)
## W = 0.96095, p-value = 0.05234
## 
## 
## $qq.resid
## 60 52 
## 59 51 
## 
## $shapiro.ID
## 
##  Shapiro-Wilk normality test
## 
## data:  coef(model)$ID[, 1]
## W = 0.97072, p-value = 0.8937
## 
## 
## $`shapiro.site:ID`
## 
##  Shapiro-Wilk normality test
## 
## data:  coef(model)$`site:ID`[, 1]
## W = 0.92175, p-value = 0.02627
## 
## 
## $qq.ranef.ID
## [1] 1 9
## 
## $`qq.ranef.site:ID`
## [1] 13 27
## 
## $res.v.fit